Skip to content

Azure: setup MPI strong scaling benchmark #8

Description

@ggorman

Azure benchmark script:

  • Script should launch job using Azure batch for different number of VM's.
  • VM's should have infiniband + high-end CPU's
  • tune
  • Rerun each experiment 3 times. Collect statistics (GFLOPS and grid-puts-per-sec).

Activity

  1. mloubout commented on Jan 24, 2020

    @mloubout
    Contributor

    Initial setup
    https://github.com/mloubout/Azure-devito

    Will be moved to devitocodes once cleanup/more ready

    To be used for devitocodes/devito#1089 too

  2. georgebisbas commented on Jan 27, 2020

    @georgebisbas

    TODO: Add a script similar to make-pbs.py in order to create job files.
    Edit update: https://github.com/devitocodes/Azure-devito/tree/azure-enrich

  3. mloubout commented on Jan 27, 2020

    @mloubout
    Contributor

    Add a script similar to make-pbs.py in order to create job files.

    You may wanna look at (but the .py way works too):
    https://batch-shipyard.readthedocs.io/en/latest/35-batch-shipyard-task-factory-merge-task/

  4. georgebisbas commented on Jan 27, 2020

    @georgebisbas

    Azure benchmark script:

    * Script should launch job using Azure batch for different number of VM's.
    

    Script exists, to be automatically generated.

    * VM's should have infiniband + high-end CPU's
    

    H* series machines

    * tune
    
    * Rerun each experiment 3 times. Collect statistics (GFLOPS and grid-puts-per-sec).
    

    It is like this by default.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions